The objective of this study was to develop a draft bear genome assembly and generate a large number of single nucleotide polymorphisms (SNP) that can be used to characterize and quantify genetic relationships among polar bears and other bear species. Six DNA pools were made from individual polar, black and brown bears (two to four individuals in each pool) from different locations in Alaska. DNA libraries were sequenced at a 6X coverage using the Illumina HiSeq sequencer. A total of one billion 100bp paired-end (PE) reads were obtained from the six libraries. Sequence reads were assembled using two strategies: “de novo” assembly and “reference assembly” using the panda and the dog genome as references. The “de novo” assembly generated approximately 90,000 contigs and 95% of the sequencing reads were mapped to those contigs. Using the reference assembly strategy, 80% of the sequencing reads were mapped to the panda genome, whereas 50% of the reads were mapped to the annotated dog genome. In total, 2.5 million SNP were detected in the bear nuclear and mitochondrial genome using this approach. 229,009 polar bear sequencing reads aligned to the dog mitochondrial genome (Broad/canFam2), 17 mitochondrial genes were clearly annotated and 154 SNP were filtered as true SNP. SNP in candidate genes and target genomic regions will be used to develop a genotyping platform to characterize genetic relationships among bear species.